{smcl}
{ul on}{bf:Title}{ul off}

   {bf:[ ] arsimsens}  {c -}{c -}  Simultaneous sensitivity analysis for Hodges-Lehmann aligned rank test

{ul on}{bf:Syntax}{ul off}

   {bf:arsimsens }{it:response} [{it:if}] [{it:in}]{bf:, {ul on}t{ul off}reat(}{it:treatment}{bf:)} {bf:{ul on}s{ul off}tr(}{it:stratum}{bf:)} [{bf:{ul on}g{ul off}am(}#{bf:)} {bf:{ul on}d{ul off}el(}#{bf:)} 
   {bf:{ul on}h{ul off}lestimate {ul on}v{ul off}erbose}]


{ul on}{bf:Description}{ul off}

   {bf:arsimsens} performs a simultaneous sensitivity analysis for the Hodges-Lehmann 
   aligned rank test of the null hypothesis after full matching, using the method 
   described in Small, Gastwirth, Kreiger, and Rosenbaum (2009).

   Specifically, it returns the maximum p-value for the Hodges-Lehmann aligned rank
   test, under unobserved confounding, for given values of gamma and delta.
   Optionally, it gives the minimum Hodges-Lehmann estimate of an additive effect, 
   for given values of gamma and delta.


{ul on}{bf:Required inputs}{ul off}

   {bf:response} : the outcome variable.

   {bf:treat(}{it:treatment}{bf:)} : a 0/1 variable indicating whether an observation 
   received the treatment (=1) or was a control (=0).
   
   {bf:str(}{it:stratum}{bf:)} : a numeric variable indicating the observation's stratum. 
   A stratum cannot have more than one of {it:both} treated and controls.
   
   
{ul on}{bf:Optional inputs}{ul off}

   {bf:gamma(}{it:#}{bf:)} : the value of Gamma for the hypothesized unobserved confounder.
   Note that this is "capital gamma," with Gamma=exp(gamma). 
   The default is 1, indicating no confounding relationship.
   
   {bf:delta(}{it:#}{bf:)} : the value of Delta for the hypothesized unobserved confounder.
   Note that this is "capital delta," with Delta=exp(delta). 
   The default is 1, indicating no confounding relationship.
   
   {bf:verbose} : display the progress of the program, by stratum.  
   
   {bf:hlestimate} : give the mimimum Hodges-Lehmann estimate for an additive effect.


{ul on}{bf:Example}{ul off}

   {bf:arsimsens myresponse, treat(mytreated) str(mygroup) gam(1.5) del(1.5) verbose}
   

{ul on}{bf:Reference}{ul off}

   Dylan Small, Joseph L. Gastwirth, Abba M. Kreiger, and Paul R. Rosenbaum. (2009). 
   "Simultaneous sensitivity analysis for observational studies using full matching 
   or matching with multiple controls." {it: Statistics and its Interface} 2: 203-211.

   
{ul on}{bf:Note}{ul off}   
	
   {bf:arsimsens} requires Stata 9 or newer.

   
{ul on}{bf:Author}{ul off}

   Daniel Lempert
   lemperds@potsdam.edu

